1. Field of the Invention
The present invention relates to a user support technique, particularly to a user support system that supports users' processes such as operations and information retrieval using agents.
2. Description of the Related Art
With the widespread use of the Internet and mobile phones, more and more people are using personal computers and various types of information terminals. In recent years, sales of electronic equipment such as personal computers have been increasing dramatically, and as a result, the number of beginners or persons who lack computer literacy has risen rapidly. With the remarkable improvements in CPU power, memory capacity and graphical user interfaces (GUIs), novices have been provided with methods that allow them to operate their equipment easily. Without aids like the GUI, novices would not be able to utilize many of the advanced functions their equipment can perform or access necessary information property.
Agent technology has become known and accepted as one form of user support technology. An agent is a personified character that appears on a computer display, obtains a user's command, and relays that command to the computer. The resulting computer output is then presented to the user as a response from the agent. (All user inputs are referred to as commands and all agent outputs are referred to as responses throughout.) The agent's presence can eliminate a great deal of awkwardness for the inexperienced user and give the impression that the computer is carrying on a conversation.
However, it requires a substantial amount of effort for a system administrator to properly manage and refine an agent system. The range of a human users commands is almost infinite. And while it is impossible to anticipate all of these commands, a user may expect the agent to accurately interpret any command he/she might give and provide a quick and proper response. Even if a substantial number of potential user commands are anticipated, it is very difficult for an agent system to search and identify the user's commands and provide near real-time response. Moreover, as the number of anticipated commands increases, the load on the server controlling the agent also increases, as it must handle a larger amount of data.